Texture Evolution and In Situ Investigation of Recrystallization Behavior in a Hot-Rolled Al-Zn-Mg-Cu-Zr Alloy

Abstract

By means of characterization techniques such as XRD, TEM, and in situ EBSD, the texture evolution, recrystallization behavior, and their modulation by the Al₃Zr phase in hot-rolled Al-Zn-Mg-Cu-Zr alloys with varied homogenization treatments were investigated. Results show that both the single-stage homogenized (SH) alloy and the double-stage homogenized (DH) alloy acquired a typical β-fiber texture after hot rolling, including brass, S, and copper orientations. The DH alloy experienced suppressed recrystallization (a recrystallization fraction of 6.05%) owing to its higher density of Al₃Zr precipitates. In contrast, the SH alloy exhibited more significant dissolution and agglomeration of Al₃Zr, leading to extensive recrystallization peaking at 78.1%. The primary recrystallization mode was identified as continuous recrystallization, characterized by the growth and coarsening of subgrains. Although dynamically recrystallized (DRx) grains formed during hot rolling could act as potential recrystallization nuclei, most of them exhibited weak growth capability, except the cube-oriented grains. During recrystallization, deformed grains with S orientation tended to transform into cube-oriented grains, while those with brass orientation prefer to convert into Goss-oriented grains. This can be attributed to the presence of highly mobile grain boundaries between these specific orientation pairs. In the DH alloy, subgrain growth and DRx grain consumption during annealing reduced orientation dispersion in deformed grains, promoting marked brass texture strengthening, with its volume fraction reaching 57.7%.

1. Introduction

Due to high specific strength and good formability, Al-Zn-Mg-Cu alloys have been widely used as a key structural material for aerospace applications [1,2,3]. However, achieving an optimal balance between their exceptional strength and other critical properties, such as corrosion resistance and fatigue performance, remains a significant challenge. The grain structure or grain boundary properties of the alloy can affect these properties. For example, recrystallized grain boundaries with large angles promote the formation of a continuous aging precipitation structure with wide precipitate-free zones (PFZs) in the Al-Zn-Mg-Cu alloy, which is detrimental to its corrosion resistance [4,5] and fatigue resistance [6,7]. By adding elements, such as Zr and Sc, to refine the alloy grains and inhibit the recrystallization process, the alloy tends to form a discontinuous aging precipitation structure along the refined grain boundaries or the subgrain boundaries during the aging process [8,9,10]. This hinders the rapid anodic corrosion along the grain boundary and significantly improves corrosion resistance of the alloy. Simultaneously, fine-grained or subgrain structures promote the formation of narrow precipitate-free zones [11,12], which can exert a favorable influence on the improvement in the alloy’s fatigue performance. In addition, following deformation processing and heat treatment, Al-Zn-Mg-Cu alloys develop distinct texture orientations, which result in significant material anisotropy and ultimately affect their service performance [13,14]. Hence, regulating the grain structure and texture compositions of the alloy plays an important role in enhancing its overall properties.

It is well known that the final grain structure of wrought aluminum alloys is determined by their initial deformation processing and subsequent heat-treatment processes. During these processes, the alloy undergoes a transformation from a deformed structure to a recrystallized structure, accompanied by texture evolution. The formation of deformation textures in aluminum alloys is influenced by factors such as the alloy’s deformation processing, composition, and secondary phases. For aluminum alloys subjected to rolling deformation, a β-fiber texture dominated by brass, copper, and S components typically forms after deformation [15,16,17]. Contrepois [16] and Zhao [18] found that brass texture is more readily formed than the copper and S textures under conditions of high temperature and large deformation in Al-Zn-Mg-Cu and Al-Cu alloys. They attributed this to the activation of additional slip systems at elevated temperatures together with the inherently low substructure energy density of the brass orientation, which collectively promote the development of this texture. Maurice [19] also found that the hot-rolling deformation at high temperatures promotes the formation of brass texture in Al-Mn-Mg and Al-Zn-Mg-Cu alloys due to the activation of the non-octahedral {112}<110> slip system. In addition to the deformation processing parameters, secondary phases also have been confirmed to have important impacts on the texture development. Zhao [20,21] revealed that an increased presence of S (Al₂CuMg) and Al₃Zr phases facilitated the formation of the brass texture in Al-Cu-Mg alloys. The S phase can induce grain rotation toward the brass orientation [20], while Al₃Zr phase suppresses dynamic recrystallization, thereby stabilizing the intensity of the brass texture [21]. It is well known that Al₃Zr particles with an L1₂ structure exhibit strong dislocation-pinning capability and inhibit recrystallization in aluminum alloys [22,23]. However, their specific influence on rolling texture deformation in Al-Zn-Mg-Cu alloys has not been fully elucidated.

The recrystallization process and accompanying texture evolution in aluminum alloys warrant considerable attention, as the final service microstructure of many alloys is typically in a recrystallized state, particularly for the heat-treatable strengthening alloys. The initial as-deformed microstructure, heat-treatment parameters, alloy composition, and secondary phases collectively govern the recrystallization kinetics and the resulting texture development. Regarding texture evolution during recrystallization, studies [24,25,26,27,28,29] have confirmed that different initial deformation textures develop into distinct recrystallization textures. Specifically, the S texture readily transforms into the cube texture because the unique orientation relationship between S and cube grains generates high-mobility Σ7 coincidence site lattice (CSL) boundaries, thereby promoting the rapid growth of cube grains [24,25,26]. In contrast, it is much more difficult for cube grains to grow within brass and copper deformation bands compared to within S deformation bands [26]. The brass texture has been observed in some Al-Cu series alloys to readily transform into the Goss texture when subjected to appropriate intermediate-temperature annealing [27,28,29]. The high-mobility Σ9 CSL boundary between brass and Goss grains facilitates this texture transformation process. Some secondary phases, such as Al₃Zr [22,23] and Al₃Sc [30], can suppress the recrystallization process and influence the grain structure of the alloy after heat treatment. Deng [31] found that the Al₃(Sc, Zr) particles formed by the combined addition of Sc and Zr significantly suppressed the recrystallization process of the Al-Zn-Mg alloy. Moreover, the brass texture developed during cold rolling was enhanced during annealing [31], However, the underlying strengthening mechanism remains unclear.

In this work, under the influence of dispersed Al₃Zr phases, a pronounced brass texture developed in the Al-Zn-Mg-Cu alloy during hot rolling and was further enhanced during subsequent annealing. The mechanism by which Al₃Zr particles affect the recrystallization behavior and texture evolution of the alloy during annealing were investigated using in situ heating EBSD experiments.

2. Experimental Procedures

The material composition used for the investigation was Al-8.9Zn-2.3Mg-1.8Cu-0.15Zr (in wt.%), and the alloy ingot was fabricated via semi-continuous casting with a pouring temperature of 710 °C and a casting speed of 1.5 mm/s. Circular ingots with a diameter of 125 mm were finally fabricated. The ingot was sectioned into blocks with a thickness of 50 mm, which were subsequently subjected to single-stage homogenization (SH, 465 °C/24 h) and double-stage homogenization (DH, 330 °C/8 h + 465 °C/18 h) treatments, respectively. Then, the homogenized alloy blocks were warmed at 450 °C for 1 h and subsequently hot-rolled to thicknesses of 5 mm (90% rolling reduction) and 2 mm (96% rolling reduction), respectively. The obtained hot-rolled sheets were annealed at 370 °C for varying durations. Samples were extracted from both the as-rolled and annealed sheets for texture measurement and microstructural characterization.

The texture compositions at the central thickness layer of the sheet samples were measured using X-ray diffraction. The XRD equipment employed was a Rigaku SmartLab polycrystalline X-ray diffractometer (Rigaku, Tokyo, Japan). The incomplete (111), (200), and (220) pole figures were determined by the Schulz back-reflection method with CuK_α radiation. Based on the incomplete pole figure data, the orientation distribution functions (ODFs) were calculated using the series expansion method with a maximum expansion order of l_max = 16. The orientation distribution functions were presented in the form of constant φ2 sections, where iso-intensity contours were plotted in Euler space defined by the Euler angles φ1, Φ, and φ2. From the ODFs, the texture volume fractions were calculated based on the method proposed by Tang [32]. If one orientation was within 15° deviation from its defined ideal component, then this orientation was classified for that specific texture.

The microstructural observations of the naturally aged samples and corresponding selected area electron diffraction (SAED) analysis were conducted using a JEOL JEM-2100F field emission transmission electron microscope (TEM) (JEOL, Tokyo, Japan) with an operating voltage of 200 kV. The samples for the electron backscatter diffraction (EBSD) experiments were prepared by mechanical polishing followed by ion beam polishing. EBSD experiments were performed on a Zeiss GeminiSEM 300 field emission scanning electron microscopy (SEM) (Carl Zeiss AG, Oberkochen, Germany) with an accelerating voltage of 20 kV. With the assistance of an in situ heating stage, the in situ EBSD experiments were conducted on the same Zeiss SEM. The samples for the in situ experiments were two hot-rolled samples that had undergone single-stage and double-stage homogenization treatments, respectively. The heating profile for the samples is shown in Figure 1. EBSD measurements were performed at four designated points on both samples. The first measurement point was at room temperature. The sample was then heated to 370 °C, where in situ measurements were performed at 1200 s and 3600 s, respectively. The acquisition time for each map was approximately 600 s. Subsequently, the sample temperature was raised to 470 °C, and a final in situ measurement was conducted after holding at this temperature for 2200 s (corresponding to 7200 s from the start of heating).

EBSD experiment samples were prepared by mechanical polish and subsequent ion beam polishing. All the EBSD measurements were performed at the mid-thickness of the samples to avoid the interference of inhomogeneous deformation along the thickness direction with the comparative analysis of the results. The EBSD analysis was performed using a step size of 0.65–1 μm, with consistent step sizes maintained across the same sample subjected to different annealing times. A confidence index threshold of ≥0.1 was applied, and the indexing rate of the deformed samples was ≥70%, while that of the annealed samples was ≥90%. All the EBSD data were analyzed by TSL OIM Analysis 7.3 (EDAX Inc., UT, USA) software. Prior to data analysis, the raw data underwent a cleanup procedure using grain dilation, with parameters set to a grain tolerance angle of 2° and a minimum grain size of 2 μm. Using this analysis soft, inverse pole figure (IPF) maps, grain size distribution maps, misorientation angle distribution maps, grain orientation spread (GOS) maps, kernel average misorientation (KAM) maps, texture component distribution maps, and phase distribution maps can be acquired and utilized to investigate the recrystallization behavior and texture evolution mechanism of the Al-Zn-Mg-Cu alloy.

3. Results

3.1. Texture Development

Figure 2 presents representative ODF sections of the Al-Zn-Mg-Cu alloys hot rolled with different reductions at 450 °C after various homogenization treatments. Figure 3 exhibits the corresponding texture orientation density distributions of the hot-rolled sheets along the α- and β-fibers.

Table 1. Volume fractions (%) of main textures for hot-rolled Al-Zn-Mg-Cu alloy sheets.

Sheets	Brass	S	Copper	Goss	Cube	Random
SH alloy (90% rolling red.)	9.7	39.9	20.4	1.8	5.1	23.1
DH alloy (90% rolling red.)	8.0	26.3	19.9	4.2	1.4	40.2
SH alloy (96% rolling red.)	14.8	21.1	20.1	0.4	2.9	40.7
DH alloy (96% rolling red.)	15.2	23.8	17.4	0.3	1.3	42

lists the volume fractions of the main textures in different hot-rolled sheets. It should be noted that the fiber connecting the copper, S, and brass textures is termed the β-fiber, while the fiber linking orientations such as Goss, brass, P, and L is referred to as the α-fiber. When the rolling reduction reached 90%, the hot-rolling textures of both the single-stage homogenized (SH) alloy and the double-stage homogenized (DH) alloy primarily consisted of copper, S, and near-Brass texture components. The strongest textures were both located in the φ₂ = 70° section, corresponding to a near-S texture with Euler angles of (55°, 35°, 70°), as shown in Figure 2a,b. The volume fraction of the S texture was significantly higher than those of the brass and copper textures, as listed in Table 1. When the rolling reduction was increased to 96%, strong brass textures developed in addition to the copper and S textures in both homogenized alloys. Although the strongest textures remained the near-S orientations with Euler angles of (55°, 75°, 30°) or (55°, 35°, 70°), their intensities were noticeably reduced. As shown in Figure 3b, with increasing rolling reduction, the intensity of the brass and copper textures increased while the near-S texture weakened accordingly, with the brass texture showing a more pronounced increment. The volume fractions of the brass texture in the SH and DH alloys increased from 9.7% and 8.0% to 14.8% and 15.2%, respectively. Notably, at a rolling reduction of 90%, the hot-rolled sheet of the SH alloy exhibited a higher level of overall texture intensity. When the rolling reduction was further increased to 96%, the texture components and their corresponding intensities of the two alloys tended to converge.

Figure 4 presents the representative ODF sections illustrating the textures of Al-Zn-Mg-Cu alloys annealed at 370 °C for different times after hot rolling with a 96% reduction. Figure 5 shows the corresponding texture orientation density distributions along the α- and β-fibers for the annealed sheets, from which the changes in intensity of the different texture components can be directly discerned.

Table 2. Volume fractions (%) of main textures for Al-Zn-Mg-Cu alloys annealed at 370 °C after hot rolling with 96% red.

Sheets	Brass	S	Copper	Goss	Cube	Random
SH alloy annealed for 1 h	35.5	16.3	12.2	14.6	8.8	12.6
DH alloy annealed for 1 h	47.2	8.6	16.3	7.4	2.3	18.2
SH alloy annealed for 4 h	14.9	5.9	9.9	3.9	61.6	3.8
DH alloy annealed for 4 h	57.7	13.5	20.8	2.6	1.5	3.9

lists the volume fractions of main textures in different annealed sheets. When annealed at 370 °C for 1 h, the brass texture intensity in the SH alloy sheet increased markedly, while the S (near-S) and copper textures weakened, accompanied by the notable formation of Goss and cube textures, as shown in Figure 4a. At this point, the volume fraction of the brass texture in the DH alloy increased from 14.8% in the as-hot-rolled state to 35.5% in the annealed state. In contrast, the DH alloy sheet exhibited a more pronounced enhancement in brass texture intensity, a slight strengthening of the copper texture, minimal change in S texture intensity, and the formation of relatively weaker recrystallization textures containing Goss and cube textures, as shown in Figure 4b. At this point, the volume fraction of the brass texture in the DH alloy increased from 15.2% in the as-hot-rolled state to 47.2% in the annealed state. When annealed at 370 °C for 4 h, the hot-rolling texture components in the SH alloy, including brass, S, and copper textures, were significantly weakened and replaced by a strong cube texture, as shown in Figure 4c, indicating that pronounced recrystallization occurred in this sheet. At this point, the volume fraction of the cube texture reached 61.6%. By comparison, after the 4 h annealing, the hot-rolling texture components in the DH alloy were all enhanced, with the brass texture showing the most pronounced increase and reaching an orientation density of 41.352, as shown in Figure 4d. At this point, the volume fraction of the brass texture increased to 57.7%. However, the intensities of the Goss and cube textures remained relatively low. This indicates that the recrystallization process was suppressed in this hot-rolled sheet. Evidently, the homogenization treatment exerts a significant influence on the recrystallization behavior and the development of annealing textures in the Al-Zn-Mg-Cu alloy.

3.2. Microstructure

Figure 6 presents TEM micrographs and SAED patterns of the hot rolled and annealed Al-Zn-Mg-Cu alloys with 96% reduction. A distinct recovery structure existed within grain interiors of the SH alloy’s hot-rolled microstructure, with the most direct evidence being polygonal dislocation structures with well-defined boundaries, as shown in Figure 6a. This indicates that a relatively obvious dynamic recovery (DRv) occurred during the hot-rolling process of the alloy. The SAED patterns of both the SH and DH alloys exhibited superlattice characteristics, indicating the formation of nanoscale coherent Al₃Zr. The Al₃Zr particles in the SH alloy had a relatively larger size and a smaller number, while those in the DH alloy were much finer and present in a much larger number. As a result, the superlattice spots of Al3Zr in the DH alloy were much clearer. It is well known that Al₃Zr particles could form in a fine, dispersed manner during the first low-temperature stage of double-stage homogenization [22,23]. Due to the pronounced pinning effect of Al₃Zr particles, the dislocation entanglement was pronounced in the DH alloy. Nevertheless, some dislocations were still able to break free from the pinning effect and form DRv substructures. After annealing at 370 °C for 4 h, the SH alloy primarily consisted of recrystallized grains with high-angle grain boundaries (HAGBs), as shown in Figure 6c, which is consistent with the predominance of recrystallized cube texture in the alloy shown in Figure 4c. As seen in Figure 6d, the annealed DH alloy mainly exhibited a substructure with low-angle grain boundaries (LAGBs), indicating that recrystallization was incomplete in this alloy. The high density of Al₃Zr particles also suppressed the recrystallization process in the Al-Zn-Mg-Cu alloy. Notably, some unclosed subgrain boundaries tended to vanish (indicated by arrows), as shown in Figure 6d, suggesting that subgrains tend to grow through coalescence with each other.

Figure 7 shows the EBSD analysis results of the Al-Zn-Mg-Cu alloys hot rolled with 96% reduction. The inverse pole figure (IPF) and texture distribution maps exhibit flattened and elongated grains in both alloys after hot rolling. The orientations of these grains are predominantly β-fiber texture components, such as brass, S, and copper, with only a small amount of Goss and very limited cube components, as shown in Figure 7a,b,d,e. These findings are consistent with the macro-texture measurement results presented in Figure 2. It must be noted that when generating the texture component distribution maps, an angular tolerance of 15° was adopted for the textures. It is worthwhile to note that Goss-oriented grains tended to form in the vicinity of brass-oriented grains, as illustrated in Figure 7b,e. Figure 7c,f shows the kernel average misorientation (KAM) maps of the as-hot-rolled microstructures of the two alloys. In EBSD, kernel average misorientation (KAM) quantifies intragranular lattice distortion by calculating the average minimum misorientation angle between a central pixel and its valid neighboring pixels within the same grain. It involves defining a kernel (first-order preferred), filtering neighbors via same-grain, low-angle threshold (5°) and quality criteria, computing pairwise minimum misorientation, and then using the arithmetic mean [33]. The KAM value is highly correlated with dislocation density and local plastic strain. A higher KAM value indicates more severe lattice distortion, typically corresponding to high dislocation density or strain concentration. It can be observed that the KAM map of DH alloy tended to display warmer colors, whereas that of SH alloy exhibits cooler colors, implying a higher average KAM value in the former. As already indicated in Figure 6, an apparent recovery occurred during hot rolling in the SH alloy. The recovery process can lead to a reduction in dislocation density, whereas in the DH alloy, dislocations were pinned and the recovery level was relatively lower. Consequently, the DH alloy retained a higher dislocation density after hot rolling, which corresponded to its higher KAM values.

Figure 8 shows the EBSD analysis results of Al-Zn-Mg-Cu alloys annealed at 370 °C for 4 h after hot rolling with 96% reduction. From the IPF and texture distribution maps, it can be observed that the SH alloy primarily consisted of near-equiaxed grains. Their grain orientations include typical recrystallization texture components such as cube and Goss, indicating a relatively high degree of recrystallization in this alloy, as shown in Figure 8a,b. In comparison, the grains in the DH alloy mostly maintained an elongated morphology, with only a small number of near-equiaxed grains (indicated by arrows). The primary texture component is the β-fiber deformation texture, suggesting a lower degree of recrystallization, as shown in Figure 8d,e. It is notable that most of the near-equiaxed grains had random texture orientations, besides a small amount of cube orientation. Figure 8c,f presents the grain orientation spread (GOS) maps for the two annealed alloys. Grain orientation spread (GOS) is defined as the average orientation deviation between each measurement point within a grain and the grain’s average orientation [34,35]. It is used to characterize the degree of strain or lattice distortion within a grain. Recrystallized grains, characterized by low dislocation density, are associated with low GOS values. The precise GOS cutoff of recrystallized grain for aluminum alloys, however, varies in the literature, where thresholds of 1° [35], 1.8° [36], and 2° [37,38] have been applied. In this study, grains with a grain orientation spread (GOS) of less than 1.3° were defined as recrystallized grains, those with a GOS ranging from 1.3° to 6° as substructured grains, and those with a GOS exceeding 6° as deformed grains. It can be seen that the SH alloy was predominantly composed of low-GOS grains. The grains with high GOS values were mainly those with brass orientation (marked by circles), which were not recrystallized. Although the SH alloy also contains some copper and S orientation grains, their GOS values were relatively lower compared to the brass-oriented grains. This implies that brass-oriented grains retained a higher level of strain, making them more resistant to recrystallization compared to other deformation texture components. This phenomenon is consistent with the result reported in the prior study [26]. In the DH alloy, apart from a small number of recrystallized grains with low GOS values, most grains exhibited high GOS values, indicating that the recrystallization process was significantly suppressed.

3.3. In Situ EBSD Investigation

Figure 9, Figure 10, Figure 11 and Figure 12 present the in situ EBSD results for the annealing process of the SH alloy subjected to a 96% hot-rolling reduction, showing IPF, texture component distribution, KAM, and second-phase distribution maps, as well as quantitative analyses of grain size and recrystallization fraction at different annealing times. It should be noted that 0 s corresponds to the as-hot-rolled state, 1200 s and 3600 s correspond to the annealed states after heating at 370 °C for the respective durations, and 7200 s corresponds to the annealed state after heating from 370 °C to 470 °C and holding for 2200 s.

From the IPF maps in Figure 9, it can be observed that the initial hot-rolled microstructure primarily consisted of deformed grains containing a high density of LAGBs (grain boundary misorientation < 15°) as shown in Figure 9a. This should be attributed to its high-density dislocation substructures formed by dynamic recovery during hot rolling. In addition, the formation of a considerable number of small grains with HAGBs (grain boundary misorientation ≥ 15°) suggests that noticeable dynamic recrystallization (DRx) occurred during the hot-rolling process (marked by circles). After heating at 370 °C for 1200 s, a significant number of near-equiaxed recrystallized grains with relatively large sizes formed. The grains exhibited a bimodal size distribution, consisting of larger grains and significantly smaller ones (circled in Figure 9b). The larger near-equiaxed grains retained only a small number of LAGBs internally. The smaller near-equiaxed grains were primarily located at the junctions of the larger grains. The simultaneous appearance of recrystallized grains with markedly different sizes after the same annealing time suggests that they should form via different recrystallization mechanisms. Furthermore, it is noteworthy that some regions still consisted of numerous subgrains (boxed in Figure 9b), indicating that the recrystallization process in these regions was suppressed. As the annealing time extended to 3600 s, the grain structure of the alloy shows no significant change, as seen in Figure 9c. When the annealing temperature is increased to 470 °C, the unrecrystallized regions diminish, and the size of the previously formed small recrystallized grains increases, as shown in Figure 9d.

Figure 10 shows the grain size distribution, misorientation angle distribution, GOS distribution, and recrystallization fraction of the SH alloy annealed for different times. As shown in Figure 10a, the grain size distribution of the deformed alloy exhibited a bimodal-like feature, corresponding to dynamically recrystallized (DRx) grains and unrecrystallized grains, respectively. With the progress of annealing, the number of small-sized grains decreased, and new grains with a size of 20–30 μm emerged, which correspond to new grains formed by static recrystallization. When the annealing temperature increased to 470 °C, the small-sized grains further decreased, and the overall grain size increased, indicating grain growth, as shown in Figure 10c,d. As shown in Figure 10e–h, the as-deformed alloy was dominated by LAGBs, and the proportion of HAGBs increased significantly after annealing. Nevertheless, since recrystallization did not occur in some regions during the entire annealing process, the alloy retained a certain proportion of LAGBs with the prolongation of annealing. Figure 10i–l presents the variations in the GOS distribution of the SH alloy at different annealing stages. After annealing, the GOS values of the alloy were mainly concentrated in the lower range, and the presence of moderate GOS distributions indicated that some substructures failed to undergo complete recrystallization. Figure 10m–p shows the proportions of recrystallized, substructured, and deformed grains at different annealing stages. In the as-deformed state, the proportions of recrystallized, substructured, and deformed grains were 39.2%, 57.1%, and 3.7%, respectively. After 1200 s of annealing, the proportion of recrystallized grains increased to 70.5%, with that of substructured grains at 29.5%. With the progression of annealing, the recrystallization proportion increased slightly and reached 78.1% after annealing at 470 °C. The remaining substructured grains with the ratio of 21.9% corresponded to the unrecrystallized regions in Figure 9d.

Figure 11 shows the maps of texture component distribution in the SH alloy corresponding to different annealing times. In the as-hot-rolled state, the primary texture components of the alloy were brass, S, and copper, while the DRx grains formed during hot rolling were predominantly cube, Goss, and randomly oriented grains (marked by circles). Regions A~I, with various texture components, are indicated in Figure 11a. Regions A and B correspond to the copper orientation; C, D, G, and I to the S orientation; Region E primarily to the two orientations containing random and cube orientation; and Regions F and H to the brass orientation. The changes in texture orientation for these regions between the initial state and after annealing at 370 °C for 1200 s are summarized in

Table 3. Texture component changes in different regions of the SH alloy before and after annealing at 370 °C.

Regions	Before Annealing	After Annealing
A	Copper	Copper
B	Copper	Cube
C	S	S
D	S	Cube
E	Random and cube	Random and cube (enlarged)
F	Brass	Goss
G	S	Random
H	Brass	Brass
I	S	Cube

. As seen in Figure 11b, Region A retained its copper orientation, while Region B transformed into a cube orientation. Notably, the copper-oriented grains in Region A grew significantly by consuming adjacent brass-oriented grains (indicated by arrow 1). Region C maintained its S orientation, whereas Region D underwent recrystallization, transforming into the cube orientation. Region G transformed into a random orientation after annealing. In Region I, only a small fraction of the original S-oriented grains remained, as they were surrounded by the recrystallized cube-oriented grain. Evidently, S-oriented grains tended to transform into cube-oriented grains during recrystallization, which aligns with previous suggestions [24,25] that cube grains preferentially grow within S-oriented deformation bands. Interestingly, Region E exhibited a mixture of cube and random orientations that were interwoven after annealing. Given the pre-existing small cube-oriented grains in this region, the newly formed cube grains likely grew from these grains. The brass-oriented grains in Region F transformed into Goss-oriented grains during annealing. The transformation of brass-oriented grains into the Goss texture during recrystallization has also been reported in previous studies on texture evolution in Al-Cu series alloys [27,28,29]. Region H, however, retained its brass texture orientation. After annealing at 370 °C for 3600 s, the grain morphology and orientation distribution in the various regions showed no significant changes. However, after holding at 470 °C, some grains underwent growth, as shown in Figure 11c,d. For instance, brass-oriented grains consumed adjacent copper, S, and Goss-oriented grains (arrows 2, 3, 4); random-oriented grains consumed adjacent brass, S, and Goss-oriented grains (arrows 5, 6, 7); and cube-oriented grains consumed surrounding S and brass-oriented grains (arrows 8, 9). These observations indicate that recrystallized grains with random and cube orientations should possess relatively stronger growth capabilities at the elevated temperature.

Figure 12 shows the KAM maps and the phase distribution maps of Al₃Zr and MgZn₂ for the SH alloy at different annealing times. As shown in Figure 12a–d, the hot-rolled SH alloy exhibited a generally high overall KAM value, which decreased markedly after annealing. However, some localized regions (marked by boxes) retained relatively high KAM values. Furthermore, after annealing, the residual subgrain boundaries inside some grains (indicated by arrows) exhibited relatively higher lattice distortion, corresponding to elevated KAM values, as illustrated in Figure 12b. From Figure 12e–h, the distribution evolution of Al₃Zr phases before and after annealing can be clearly observed. In the as-hot-rolled state, the Al₃Zr phases in the alloy exhibited a relatively heterogeneous distribution. After annealing treatment, the number of Al₃Zr phases decreased, accompanied by an increase in distribution inhomogeneity. It can been observed that Region 1 retained a high density of Al₃Zr phase particles, whereas distinct phase segregation occurred in Regions 2 and 3 (marked by the boxes). The regions with Al₃Zr aggregation were largely consistent with the areas featuring relatively higher KAM values. Considering that the strong recrystallization-inhibiting effect of Al₃Zr particles, it is reasonable that Region 1, containing a higher density of these particles, did not undergo significant recrystallization. It is worth noting that when the annealing temperature was increased to 470 °C, the locations of Al₃Zr aggregation shifted. The number of Al₃Zr particles in Regions 2 and 3 decreased substantially, while new, smaller Al₃Zr aggregations formed in Regions 4 and 5, leading to a slight increase in KAM values in those areas. Combined with the observations in Figure 11, Grain E in Figure 12b,f developed into a fully cube-oriented grain along with the dissolution of Al₃Zr phases in Region 2. The density of Al₃Zr phases in Region 6 increased, as shown in Figure 12h. However, the KAM value of this region decreased. This suggests that the dislocation-pinning effect of Al₃Zr particles weakened at elevated temperatures, thereby enabling recrystallization to occur in this region. From Figure 12i–l, it is evident that compared with low-temperature annealing, the number of MgZn₂ particles decreased more markedly during high-temperature annealing. This reduction can be attributed to their continuous dissolution into the matrix.

Figure 13, Figure 14 and Figure 15 present the results of in situ EBSD study on the annealing process of the DH alloy with a 96% hot-rolling reduction, showing the IPF maps, texture component maps, GOS maps, KAM maps, and second-phase distribution maps, as well as quantitative analyses of grain size and recrystallization fraction at different annealing times. The testing temperatures and time points were consistent with those for the SH alloy. As can be seen from the IPF maps in Figure 13a–d, most grains retained the morphology resulting from hot-rolling deformation, indicating a low degree of recrystallization in the DH alloy during annealing. As illustrated in Figure 13a, in the initial as-hot-rolled state, Regions 1–4 (marked by boxes) contained a number of small-sized near-equiaxed grains. These small grains were primarily located at the boundaries between large deformed grains (e.g., Regions 2, 3, and 4) and should be formed by DRx during the hot-rolling process. Notably, the number of these small-sized DRx grains was significantly lower than that in the SH alloy. With the progression of annealing, the number of DRx grains decreased, and some of these grains were consumed by neighboring deformed grains. For instance, the DRx grain in Region 1 shrank until it vanished. This observation demonstrates that the DRx grains formed during hot rolling were not stable. When the annealing temperature was increased to 470 °C, the number of LAGBs inside the grains decreased, suggesting that the subgrains underwent coalescence and growth. In addition, a distinct recrystallized grain (Grain B) was formed, while other regions remained predominantly in the unrecrystallized state, as shown in Figure 13d. Figure 13e–h shows the texture component distribution maps corresponding to different annealing times. The primary textures of the sample both before and after annealing were the dominant brass texture along with the minor S and copper textures. In the as-hot-rolled state, except for a small number of cube-oriented grains, most of the small-sized DRx grains exhibited random orientations, as seen in Figure 13e. It is worth noting that Grain A simultaneously possessed brass, S, and copper components (marked by arrows), and Grain B formed during high-temperature annealing exhibited a random orientation. Figure 13i–l shows the grain orientation spread (GOS) maps corresponding to different annealing times. In the as-hot-rolled state, there were several scattered regions with low GOS values, which should correspond to the microstructures formed by dynamic recovery or recrystallization during the hot-rolling process. Notably, after annealing at 370 °C, the number of small low-GOS regions in Region 5 decreased, as shown in Figure 13h. When the annealing temperature was increased to 470 °C, the newly formed recrystallized Grain B exhibited a remarkably low GOS value.

Figure 14 shows the grain size distribution, misorientation angle distribution, GOS distribution, and recrystallization fraction of the DH alloy annealed for different times. Owing to the suppressed recrystallization of the DH alloy, the grain morphology changed slightly, and thus, the grain size distribution exhibited a minor difference before and after annealing, as shown in Figure 14a–d. For the same reason, the grain boundaries of the alloy were dominated by LAGBs both before and after annealing, as illustrated in Figure 14e–h. However, the average grain size increased after annealing, which was primarily due to the reduction in DRx grains. In comparison with the SH alloy, the DH alloy in the as-hot-rolled state had a smaller proportion of low GOS values, while the proportions of moderate and high GOS values increased significantly. With the progression of annealing, the GOS values were mainly concentrated in the ranges of 4–5° and 6–7°, and the proportion of high GOS values decreased gradually, as presented in Figure 14i–l. As shown in Figure 14m–p, the proportion of recrystallized grains in the as-deformed DH alloy was 7.0%, and it decreased to 4.9% after 1200 s of annealing, indicating that the dynamically recrystallized (DRx) grains formed during hot rolling were consumed by the surrounding non-recrystallized grains. After annealing at 470 °C, the recrystallization fraction increased slightly to 6.05%, and the decrease in the deformed grain fraction was mainly attributed to continuous recovery and the occurrence of recrystallization in localized regions.

Figure 15 shows the KAM maps and the phase distribution maps of Al₃Zr and MgZn₂ for the DH alloy at different annealing times. As shown in Figure 15a–d, the as-hot-rolled sample of the DH alloy exhibited a generally high overall KAM value. After annealing at 370 °C, the sample still maintained a high KAM level, except for a few small localized regions (indicated by arrows). Following annealing at 470 °C, the number and size of these low-KAM regions increased, as illustrated in Figure 15d. A reduction in the KAM value indicated a decrease in the dislocation density of the corresponding region. Specifically, during annealing, the growth of intragranular subgrains and the coalescence of DRx grains with adjacent grains decreased the dislocation density of these regions, thereby leading to a reduction in their KAM values. Figure 15e,f illustrates the distribution evolution of Al₃Zr phases before and after annealing. In the as-hot-rolled state, the DH alloy exhibited a higher density of Al₃Zr particles with a relatively more homogeneous distribution as compared with the SH alloy. It is apparent that the number of Al₃Zr phases decreased slightly after annealing at 370 °C, whereas a more distinct reduction was observed after annealing at 470 °C. Nevertheless, the extent of the decrease remained relatively limited. It is noteworthy that Al₃Zr particles aggregated in the newly formed recrystallized Grain B, and the corresponding KAM values in that region, also decreased, as shown in Figure 15d,h. This suggests that the aggregating process of Al₃Zr should be accompanied by a reduction in dislocation density. From Figure 15i,l, it is evident that similar to the SH alloy, the number of MgZn₂ phases decreased continuously as annealing proceeded. This reduction in the quantity of MgZn₂ phases could be attributed to their continuous dissolution into the matrix.

Based on the phase distribution maps in Figure 12 and Figure 15, the variations in phase fractions of alloys with different homogenization treatments are summarized in Figure 16. In the as-hot-rolled state, the DH alloy exhibited a higher density of Al₃Zr phases, with its area fraction reaching 22.1% compared to 8.9% in the SH alloy. After annealing, the density of Al₃Zr phases in both alloys decreased significantly; nevertheless, the DH alloy still maintained a relatively high Al₃Zr phase density, with the area fractions of Al₃Zr phases in the DH and SH alloys reaching 13.1% and 2.6%, respectively, after annealing at 470 °C. Both alloys had a low density of MgZn₂ phases, with the SH alloy showing a relatively higher density. The density of MgZn₂ phases in both alloys decreased to a low level with the progression of annealing.

4. Discussion

4.1. Hot-Rolling Deformation Behavior and Texture Development

It is well established that the rolling texture of aluminum alloys typically consists of brass, S, and copper components [15,16,17], whereas the recrystallization texture is predominantly composed of cube, ND-rotated cube, Goss, and P components [22,23,24,25,39,40]. Notably, the Goss texture can also be obtained through deformation processing [21,41,42]. The formation of hot-rolling textures in aluminum alloys is primarily influenced by rolling process parameters [16,18,19], alloy composition, and second-phase distribution [16,20,21]. Contrepois [16], in a study on texture evolution during hot rolling of an Al-Zn-Mg-Cu alloy between 420 °C and 450 °C, also found that high-temperature and large-strain conditions favored the formation of the brass texture. Contrepois [16] and Maurice [19] attributed this to the activation of the additional non-octahedral {112}<110> slip systems at elevated rolling temperatures, which promotes the stabilization and development of the brass texture. From a thermodynamic perspective, Zhao [18] demonstrated that under large-strain conditions, the brass texture with lower substructure energy density is more stable.

In the present study, the Zr-added Al-Zn-Mg-Cu alloy, subjected to different homogenization treatments, was hot-rolled at a relatively high temperature (450 °C), resulting in the formation of typical brass, S, and copper textures, as shown in Figure 2. Increasing the rolling reduction from 90% to 96% resulted in a weakening of the near-S texture and a pronounced enhancement in the brass texture, as shown in Figure 2 and Figure 3. This indicates that under high-temperature deformation conditions, greater strain facilitates the formation of the brass texture, which is consistent with the findings of aforementioned studies [16,18,19]. Zhao [21] investigated the effect of Zr addition on the hot-rolling texture of an Al-Cu-Mg alloy. The results revealed that the Al₃Zr phase inhibited the dynamic recrystallization (DRx) process of the alloy during hot rolling, which consequently suppressed the transformation from the brass texture to the Goss texture. This led to the formation of a stronger brass texture in the Al-Cu-Mg-Zr alloy. In the present study, Al₃Zr particles were also present in the alloy. However, the alloy subjected to double-stage homogenization (DH) treatment exhibited a higher density of Al₃Zr particles compared to that subjected to single-stage homogenization (SH) treatment, as shown in Figure 6, Figure 12 and Figure 15. At 90% rolling reduction, the SH alloy demonstrated an overall higher texture intensity than the DH alloy, whereas the difference in texture components between the two alloys became less pronounced at 96% rolling reduction. This difference is likely attributable to the dynamic recovery (DRv) and dynamic recrystallization (DRX) occurring during hot rolling. As shown in Figure 6, distinct DRv substructures formed in both alloys at 96% reduction. Concurrently, EBSD analysis indicated obvious DRx at the grain boundaries of the deformed grains in both alloys, as evidenced in Figure 9a and Figure 13a. Duckham [43] suggested that during hot deformation, the formation and growth of subgrains during DRv cause the scattered orientations within a grain to converge toward a stable orientation, thereby promoting the development of a strong deformation β-fiber texture. Therefore, the pronounced DRv observed in this study can facilitate the formation of strong brass, S, and copper textures. Considering its lower Al₃Zr particle density, the SH alloy was expected to experience a higher degree of DRv, which accounts for its higher overall texture intensity at 90% reduction. However, when the rolling reduction was increased to 96%, the enhanced accumulated strain effectively facilitated more substantial DRx behavior. Zhao [21] found that DRx consumes deformed grains and markedly reduces the texture intensity in Al-Cu-Mg alloys. Consequently, as DRx intensified under the condition of elevated rolling reduction, the deformation texture underwent gradual consumption. This, in turn, resulted in a overall decrease in texture intensity. Notably, the as-hot-rolled SH alloy exhibited a larger number of small-sized recrystallized grains, which indicated more pronounced DRx behavior, as illustrated in Figure 9a. This led to more extensive consumption of the deformation texture, finally resulting in a convergence in texture characteristics between the two alloys.

As shown in Figure 6, both SH and DH alloys exhibited distinct intragranular recovery substructures, indicating that pronounced DRv occurred during the hot-rolling process. However, the recovery substructures in the DH alloy were smaller in size, and a higher density of dislocation entanglements was retained. Thus, the DH alloy exhibited a lower degree of DRv and correspondingly possessed a higher dislocation density. This was primarily attributed to the strong pinning effect of high-density Al₃Zr particles on dislocation movement, suppressing the DRv process. As a result, the higher dislocation density in the DH alloy consequently resulted in a relatively elevated level of KAM distribution, as illustrated in Figure 7c,f. As mentioned earlier, the SH alloy exhibited a higher degree of DRx than the DH alloy, which is also attributed to the effect of the Al₃Zr particles. In the SH alloy, distinct DRx grains were observed both at the boundaries and within the interior of the original deformed grains, whereas in the DH alloy, DRX grains were primarily located at the boundaries of deformed grains, as shown in Figure 9a and Figure 13a. The lower density and relatively inhomogeneous distribution of Al₃Zr particles in the SH alloy led to the occurrence of significant DRx within the interior regions of deformed grains where the particle density was low. In contrast, the higher density and more uniform distribution of Al₃Zr particles in the DH alloy confined DRx mainly to the grain boundary regions, where high strain accumulated during hot rolling.

Furthermore, an interesting phenomenon is that a single deformed grain in the hot-rolled state contains multiple texture components, as illustrated in Figure 13a,e. As is well known, the grain orientation of the deformed alloy is determined by dislocation slip and the consequent grain rotation, and plastic deformation in polycrystalline materials usually activates multiple slip systems to accommodate intergranular deformation [16,19]. This indicates that different regions within a grain experienced different slip activities and grain rotation processes, resulting in different final texture orientations. However, regions with S and brass orientations, as well as those with S and copper orientations, tended to be adjacent to or intertwined with each other, whereas regions with brass and copper orientations were generally separated, as shown in Figure 11a and Figure 13e. This phenomenon can be explained by the orientation relationships among the brass, S, and copper textures. In Euler space, the β-fiber appears as a continuous curve starting from the brass orientation (approximately φ₁ = 35°, Φ = 45°, φ₂ = 90°), passing through the S orientation (approximately φ₁ = 59°, Φ = 37°, φ₂ = 65°), and terminating at the copper orientation (approximately φ₁ = 90°, Φ = 35°, φ₂ = 45°). Since the misorientation angles between the S orientation and both the brass and copper orientations are relatively small, the S orientation acts as a bridge between the brass and copper orientations. During rolling deformation, the grain orientations can evolve and transition relatively easily and continuously along the β-fiber following the path of brass ↔ S ↔ copper. Consequently, regions with S/brass and S/copper orientations tend to be adjacent to or interwoven with each other. This also explains why, as the rolling deformation increased, the intensity of the near-S texture decreased while the intensities of the brass textures were enhanced, as shown in Figure 2 and Figure 3.

4.2. The Annnealing Behavior and Texture Evolution

As is well known, deformed aluminum alloys undergo recovery or recrystallization and grain growth during annealing. Correspondingly, the grain morphology and texture orientation of the deformed microstructure also undergo distinct changes. In this study, different homogenization treatments resulted in variations in the density and distribution uniformity of Al₃Zr phase particles, leading to distinct texture and microstructural evolution during annealing in the SH and DH alloys. During annealing at 370 °C, the SH alloy exhibited the formation of pronounced recrystallization textures such as cube and Goss, whereas the DH alloy showed weaker cube and Goss textures and a significant strengthening of the brass texture, as illustrated in Figure 4 and Figure 5. As mentioned repeatedly earlier, Al₃Zr particles, which can pin dislocations, exert an inhibitory effect on recrystallization. The higher density and more uniform distribution of Al₃Zr particles in the DH alloy resulted in a more pronounced suppression of recrystallization, thereby hindering the development of cube and Goss textures. The formation of high-angle recrystallized grain boundaries in the SH alloy (Figure 6c) and the presence of numerous near-equiaxed grains (Figure 8) indicated a higher degree of recrystallization in this alloy, which, in turn, promoted the formation of Goss and cube textures.

The recrystallization modes in aluminum alloys mainly include continuous recrystallization and discontinuous recrystallization. The former is realized by the coalescence and rotation of subgrains, which gradually transform subgrain boundaries into high-angle grain boundaries, while the latter is achieved through the nucleation of new recrystallized grains and their growth via interface migration [44]. Figure 6d shows the phenomenon that some unclosed subgrain boundaries tended to disappear, indicating the occurrence of subgrain coalescence in the DH alloy. As shown in Figure 9b, the dislocation substructures with LAGBs inside the original deformed grains of the SH alloy were significantly reduced during annealing, and some residual subgrain boundaries were left within the grains (marked by arrows), which also indicates the occurrence of subgrain coalescence. Although subgrain coalescence occurs in both alloys, it does not mean that the two alloys undergo recrystallization to the same extent. As shown in Figure 13a–d, despite the occurrence of subgrain coalescence in the DH alloy, the migration of subgrain boundaries was hindered by Al₃Zr particles, preventing the subgrain boundaries from evolving into recrystallized grains. Thus, the recrystallization level of this alloy was much lower. Notably, by comparing Figure 9b and Figure 11b, it can be found that residual subgrain boundaries or subgrains were retained inside the grains with S, copper, cube, and Goss orientations, indicating that grains with these orientations can all be formed via continuous recrystallization. Considering that subgrain coalescence also occurs in the deformed grains with brass orientation (Figure 13), the large-sized brass-oriented grains without internal substructures in Figure 9 likely formed through the continuous recrystallization mechanism dominated by subgrain coalescence.

In addition to continuous recrystallization, another atypical recrystallization mechanism occurs in the Al-Zn-Mg-Cu alloy in this study. As shown in Figure 9b, small-sized recrystallized grains existed between some large-sized grains (marked by circles), and the regions where these small grains were located contained even finer DRx grains in the as-hot-rolled state. This implies that the DRx grains may directly serve as recrystallization nuclei and continue to grow during annealing. However, under this mechanism, most of the recrystallized grains exhibited weak growth capability. With the progression of annealing, the recrystallized grains in these regions were even engulfed by the surrounding continuously recrystallized grains. Only a few DRx grains with cube orientation could achieve significant growth during annealing, such as Grain E and another grain indicated by the arrow in Figure 11b. Furthermore, the continuously recrystallized grains with other orientations exhibited different growth abilities during high-temperature annealing. As shown in Figure 11, the recrystallized grains with random and cube orientations possess stronger growth capabilities in comparison with the grains with other orientations.

One noteworthy phenomenon in this study is the localized aggregation of Al₃Zr phases during annealing. As annealing proceeded, the overall quantity of Al₃Zr phases decreased, indicating its dissolution back into the matrix. The concurrent localized aggregation in specific areas suggests the reprecipitation and coarsening of Al₃Zr, as shown in Figure 12 and Figure 15. Furthermore, the locations of Al₃Zr aggregations shift during high-temperature annealing (Figure 12h), providing additional evidence that the Al₃Zr phase should undergo a dynamic process of dissolution and reprecipitation throughout annealing. Regions where no significant dissolution of Al₃Zr occurred naturally exhibited suppressed recrystallization, as exemplified by Region 1 in Figure 12b,f. In contrast, regions where Al₃Zr aggregation took place correspond to recrystallized areas, exemplified by Grain E in Figure 12 and Grain B in Figure 15. This appears to contradict the well-known recrystallization-inhibiting effect of Al₃Zr particles. However, relevant studies indicate that at elevated temperatures, fine Al₃Zr particles with the L₁₂ structure can undergo coarsening and a transition to the D0₂₃ structure [23,45]. Given that hot rolling introduces high dislocation density and strain energy, this process of coarsening and aggregation is likely accelerated. Upon transforming to the D0₂₃ structure, Al₃Zr loses coherency with the Al matrix, drastically weakening its dislocation-pinning ability and thereby its capacity to suppress recrystallization [23]. This facilitates dislocation motion and promotes recrystallization in the regions of Al₃Zr aggregation. It is notable that within the SH alloy, the KAM values in regions of Al₃Zr aggregation were higher than in other recrystallized areas. This is primarily because other recrystallized regions correspond to areas with far fewer Al₃Zr particles, where dislocation density was greatly reduced through subgrain coalescence and growth. In contrast, the aggregated Al₃Zr particles retained a capacity to impede dislocation motion, which maintained a relatively higher residual dislocation density and, in turn, led to relatively higher KAM values in these areas. Nevertheless, compared to the initial as-deformed microstructure, the overall KAM in the aggregation regions was lower, as evidenced in Figure 12 and Figure 15. Consequently, Grain B in Figure 15 exhibited a lower KAM value relative to the surrounding unrecrystallized grains. Based on the above, the evolution of Al₃Zr distribution during the annealing process and its effect on the development of recrystallization (Rx) grains are summarized in Figure 17.

The in situ annealing EBSD study revealed the transformation patterns of texture components in grains with different initial orientations during annealing, as shown in Figure 11 and Table 3. Grains with copper, S, and brass orientations can retain their original orientations via continuous recrystallization, exemplified by Grains A, C, and H in Figure 11. In contrast, grains with S and brass orientations can transform into the typical recrystallization textures of cube and Goss, corresponding to Grains D and F, respectively. Relevant studies [24,25,26] have found that S-oriented and cube-oriented grains exhibit an orientation relationship of approximately 40° about <111>axis. Grain boundaries between S and cube grains with this relationship typically form Σ7 coincidence site lattice (CSL) boundaries, which exhibit high mobility [25,26]. Consequently, S-oriented deformed grains readily transform into cube-oriented grains during recrystallization. Similarly, brass-oriented and Goss-oriented grains exhibit an orientation relationship of approximately 40° about <110>axis. Boundaries with this relationship typically correspond to Σ9 CSL boundaries, which also possess high mobility [28,29]. As a result, brass-oriented deformed grains tend to transform into Goss-oriented grains during annealing. Previous studies [26] have indicated that cube-oriented nuclei can also form within copper deformation bands. However, their growth capability is significantly lower than that of nuclei formed within S deformation bands [26]. Therefore, although the copper-oriented grain in Region B transformed into a cube grain after annealing, it was more likely the result of growth from a cube nucleus that originated in the adjacent S-oriented deformed grain, as illustrated in Figure 11.

Figure 18 shows magnified views of selected local regions extracted from Figure 9b, containing representative grains with the orientations close to brass, S, cube, and Goss. Using the EBSD analysis software, the crystallographic orientation relationship between adjacent grains can be determined, as illustrated in

Table 4. Misorientation characteristics between near-S-/cube-oriented and near-brass-/Goss-oriented grains in the SH alloy.

Grains	Miller Indices	Orientation Relationship	Types of CSL Boundary
1 and 2	(6 2 9)[8 21 −10] and (0 1 5)[−1 0 0]	35.9°[13 −12 15]	Σ7
1 and 3	(6 2 9)[8 21 −10] and (1 1 17)[−3 20 −1]	44.9°[2 −3 4]	/
4 and 5	(3 18 20)[10 5 −6] and (12 2 13)[−1 32 −4]	40.7°[−2 −7 −7]	/
6 and 7	(1 6 7)[16 9 −10] and (0 3 4) [38 4 −3]	31.9°[3 −16 −13]	/

. For instance, the relationship between Grain 1 and Grain 2 is 35.9° about the axis [13 −12 15], which is close to the 40° about <111> relationship. The boundary between them is identified as a highly mobile Σ7 CSL boundary. During high-temperature annealing, the S-oriented Grain 1 was consumed by Grain 2 (seen in Figure 9d), further supporting the high mobility of their interface. However, the orientation relationship between Grain 1 and another cube-oriented grain (Grain 3) is 44.9° about [2 −3 4], which deviates more significantly from the ideal 40° about <111>. The fact that this cube-oriented grain did not consume the S-oriented grain during high-temperature annealing suggests their shared boundary had lower mobility. Grains 4 and 5, as well as Grains 6 and 7, are pairs of brass and cube-oriented grains, respectively. Their orientation relationships are close to 40° about <110>, but Σ9 CSL boundaries were not identified between them. Consequently, no mutual consumption occurred during subsequent annealing at the high temperature. Zhao [28,29] reported that distinct Σ9 boundaries exist between Goss- and brass-oriented grains in Al-Cu-Mg alloys during annealing at moderate temperatures. In this study, it can be speculated that Σ9 boundaries likely existed in the early annealing stages to promote the transformation from brass to Goss. As grain growth proceeded, the orientation relationship between them changed, gradually deviating from the ideal 40° about <110>, causing the high-mobility Σ9 boundaries to disappear. Therefore, although the grain boundaries close to Σ7 and Σ9 were identified only in a limited number of grain pairs, they still hold important reference value for investigating the origin of recrystallization and deducing the transformation laws of different textures.

For the DH alloy, a prominent feature in its texture evolution is the significant enhancement in the brass texture intensity during annealing. This phenomenon can be attributed to the behavior of its recovery process. The distinct subgrain structure observed in Figure 6d and the reduction in the number of substructures within deformed grains shown in Figure 13a–d both indicate that pronounced recovery occurred in the DH alloy. Figure 13i–l displays the changes in the GOS distribution of the DH alloy at different annealing times. It can be observed that Region 5 in the as-hot-rolled state is predominantly composed of high-GOS regions, with scattered low-GOS zones embedded therein, as shown in Figure 13i. These low-GOS domains are largely ascribed to DRx or DRv occurring during the hot-rolling process. Due to the presence of DRx grains and some DRv subgrains with relatively high-angle grain boundaries, the orientation of this region remains relatively dispersed in the as-hot-rolled state. During annealing, the static recovery, involving the formation and growth of subgrains, causes these initially dispersed orientations within Region 5 to gradually converge towards the brass orientation. As shown in Figure 13j–l, the number of low-GOS zones in Region 5 decreases, signifying that the recovery process promotes the convergence of intragranular dispersed orientations, thereby facilitating the formation of a strong brass texture.

5. Conclusions

This study systematically investigated the texture evolution laws, recrystallization behavior, and the influence of the Al₃Zr phase on microstructure formation and texture development during hot rolling and annealing of Al-Zn-Mg-Cu-Zr alloys subjected to different homogenization treatments. The main conclusions are as follows:

Both the SH and DH alloys formed typical β-fiber textures dominated by brass, S, and copper orientations after hot rolling. The texture with the maximum intensity in both alloys was the near-S texture. With rolling reduction increased from 90% to 96%, the near-S texture weakened, and brass texture intensity rose significantly in both alloys. The corresponding brass texture volume fractions in the SH and DH alloys reached 14.8% and 15.2%, respectively.

As annealing proceeded, Al₃Zr phases underwent dynamic dissolution and reprecipitation. In the SH alloy, Al₃Zr phases decreased with more pronounced aggregation during annealing, leading to recrystallization peaking at 78.1% and the formation of cube/Goss-dominated recrystallization textures. In contrast, the DH alloy retained a high density of Al₃Zr particles, which effectively suppressed recrystallization (only 6.05% fraction) and thus preserved the dominant brass orientation.

Continuous recrystallization was identified as the primary recrystallization mode in the Al-Zn-Mg-Cu-Zr alloy. Deformed grains completed continuous recrystallization through coalescence and coarsening. While the dynamically recrystallized (DRx) grains formed during hot rolling could also act as nuclei for growth, their growth capability was relatively weak, except for cube-oriented grains.

During annealing, S-oriented deformed grains tended to convert to cube-oriented grains via recrystallization, while brass-oriented ones transformed into Goss-oriented grains. As annealing time prolonged, the brass texture in the DH alloy was significantly strengthened, peaking at a volume fraction of 57.7%. This was mainly due to subgrain growth and DRx grain consumption during recovery, reducing orientation dispersion in deformed grains.